LKML Archive on lore.kernel.org help / color / mirror / Atom feed
From: Jeremy Fitzhardinge <Jeremy.Fitzhardinge@citrix.com> To: Jan Beulich <jbeulich@novell.com> Cc: Jeremy Fitzhardinge <jeremy@xensource.com>, Ingo Molnar <mingo@elte.hu>, linux-kernel@vger.kernel.org Subject: Re: [PATCH] adjust/fix LDT handling for Xen Date: Fri, 11 Jan 2008 09:28:14 -0800 [thread overview] Message-ID: <4787A72E.8060704@citrix.com> (raw) In-Reply-To: <4787436E.76E4.0078.0@novell.com> Jan Beulich wrote: > Don't rely on kmalloc(PAGE_SIZE) returning PAGE_SIZE aligned memory > (Xen requires GDT *and* LDT to be page-aligned). Can kmalloc return non-page-aligned PAGE_SIZE allocations? > Using the page > allocator interface also removes the (albeit small) slab allocator > overhead. Runtime or space overhead? Given that they're once-off allocations, the time part isn't a big factor. And apparently LDT is completely unused. Seems like a reasonable change either way. > The same change being done for 64-bits for consistency. > > Further, the Xen hypercall interface expects the LDT address to be > virtual, not machine. > Shows how much the LDT gets used then... Hm, but I didn't make that code up; I wonder where it came from? But certainly, passing the mfn of the first page of a vmalloc area doesn't make any sense. > Signed-off-by: Jan Beulich <jbeulich@novell.com> > Cc: Ingo Molnar <mingo@elte.hu> > Acked-by: Jeremy Fitzhardinge <jeremy@xensource.com> > --- > arch/x86/kernel/ldt_32.c | 7 +++---- > arch/x86/kernel/ldt_64.c | 7 +++---- > arch/x86/xen/enlighten.c | 9 +-------- > 3 files changed, 7 insertions(+), 16 deletions(-) > > --- linux-2.6.24-rc7/arch/x86/kernel/ldt_32.c 2008-01-10 16:53:54.000000000 +0100 > +++ 2.6.24-rc7-x86-xen-ldt/arch/x86/kernel/ldt_32.c 2008-01-09 13:59:50.000000000 +0100 > @@ -9,7 +9,6 @@ > #include <linux/mm.h> > #include <linux/smp.h> > #include <linux/vmalloc.h> > -#include <linux/slab.h> > > #include <asm/uaccess.h> > #include <asm/system.h> > @@ -38,7 +37,7 @@ static int alloc_ldt(mm_context_t *pc, i > if (mincount*LDT_ENTRY_SIZE > PAGE_SIZE) > newldt = vmalloc(mincount*LDT_ENTRY_SIZE); > else > - newldt = kmalloc(mincount*LDT_ENTRY_SIZE, GFP_KERNEL); > + newldt = (void *)__get_free_page(GFP_KERNEL); > > if (!newldt) > return -ENOMEM; > @@ -69,7 +68,7 @@ static int alloc_ldt(mm_context_t *pc, i > if (oldsize*LDT_ENTRY_SIZE > PAGE_SIZE) > vfree(oldldt); > else > - kfree(oldldt); > + put_page(virt_to_page(oldldt)); > } > return 0; > } > @@ -114,7 +113,7 @@ void destroy_context(struct mm_struct *m > if (mm->context.size*LDT_ENTRY_SIZE > PAGE_SIZE) > vfree(mm->context.ldt); > else > - kfree(mm->context.ldt); > + put_page(virt_to_page(mm->context.ldt)); > mm->context.size = 0; > } > } > --- linux-2.6.24-rc7/arch/x86/kernel/ldt_64.c 2008-01-10 16:53:54.000000000 +0100 > +++ 2.6.24-rc7-x86-xen-ldt/arch/x86/kernel/ldt_64.c 2008-01-09 13:59:50.000000000 +0100 > @@ -12,7 +12,6 @@ > #include <linux/mm.h> > #include <linux/smp.h> > #include <linux/vmalloc.h> > -#include <linux/slab.h> > > #include <asm/uaccess.h> > #include <asm/system.h> > @@ -41,7 +40,7 @@ static int alloc_ldt(mm_context_t *pc, u > if (mincount*LDT_ENTRY_SIZE > PAGE_SIZE) > newldt = vmalloc(mincount*LDT_ENTRY_SIZE); > else > - newldt = kmalloc(mincount*LDT_ENTRY_SIZE, GFP_KERNEL); > + newldt = (void *)__get_free_page(GFP_KERNEL); > > if (!newldt) > return -ENOMEM; > @@ -73,7 +72,7 @@ static int alloc_ldt(mm_context_t *pc, u > if (oldsize*LDT_ENTRY_SIZE > PAGE_SIZE) > vfree(oldldt); > else > - kfree(oldldt); > + put_page(virt_to_page(oldldt)); > } > return 0; > } > @@ -117,7 +116,7 @@ void destroy_context(struct mm_struct *m > if ((unsigned)mm->context.size*LDT_ENTRY_SIZE > PAGE_SIZE) > vfree(mm->context.ldt); > else > - kfree(mm->context.ldt); > + put_page(virt_to_page(mm->context.ldt)); > mm->context.size = 0; > } > } > --- linux-2.6.24-rc7/arch/x86/xen/enlighten.c 2008-01-10 16:53:55.000000000 +0100 > +++ 2.6.24-rc7-x86-xen-ldt/arch/x86/xen/enlighten.c 2008-01-09 14:00:51.000000000 +0100 > @@ -275,19 +275,12 @@ static unsigned long xen_store_tr(void) > > static void xen_set_ldt(const void *addr, unsigned entries) > { > - unsigned long linear_addr = (unsigned long)addr; > struct mmuext_op *op; > struct multicall_space mcs = xen_mc_entry(sizeof(*op)); > > op = mcs.args; > op->cmd = MMUEXT_SET_LDT; > - if (linear_addr) { > - /* ldt my be vmalloced, use arbitrary_virt_to_machine */ > - xmaddr_t maddr; > - maddr = arbitrary_virt_to_machine((unsigned long)addr); > - linear_addr = (unsigned long)maddr.maddr; > - } > - op->arg1.linear_addr = linear_addr; > + op->arg1.linear_addr = (unsigned long)addr; > op->arg2.nr_ents = entries; > > MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF); > > > >
next prev parent reply other threads:[~2008-01-11 17:29 UTC|newest] Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top 2008-01-11 9:22 [PATCH] adjust/fix LDT handling for Xen Jan Beulich 2008-01-11 9:39 ` Jiri Slaby 2008-01-11 10:08 ` Jan Beulich 2008-01-11 17:28 ` Jeremy Fitzhardinge [this message] 2008-01-14 8:22 ` Jan Beulich 2008-01-14 15:48 ` Ingo Molnar 2008-01-14 16:06 ` Jan Beulich 2008-01-14 17:04 ` Jeremy Fitzhardinge 2008-01-15 13:24 ` Ingo Molnar 2008-01-15 21:47 ` Jeremy Fitzhardinge
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=4787A72E.8060704@citrix.com \ --to=jeremy.fitzhardinge@citrix.com \ --cc=jbeulich@novell.com \ --cc=jeremy@xensource.com \ --cc=linux-kernel@vger.kernel.org \ --cc=mingo@elte.hu \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).